J. Chem. SOC., Faraday Trans. I, 1987, 83, 13-27 hv Magnetic Resonance of Ultrafast Chemical Reactions L \ < l o p s B2'P-(QFe) ca. 10ns \ ca. 200 ps 3*BzP(QFe) B2'P(QFe)- cu. 1 ms J Examples from Photosynthesis J. R. Norris, C. P. Lin and D. E. Budil Chemistry Division, Argonne National Laboratory, Argonne, Illinois 60439, U.S.A . and Department of Chemistry, University of Chicago, Chicago, Illinois 60637, U.S.A. Using e.s.e. (electron spin-echo), RYDMR (reaction-yield-detected mag- netic resonance) and e.s.r. (electron spin resonance) in protein single crystals, we have investigated the cation, the triplet and the initial radical pair associated with the process of photoinduced charge separation in photo- synthesis. The initial charge separation of bacterial photosynthesis occurs in a few picoseconds within in a reaction-centre protein containing eight electron-transfer components, namely four bacterial chlorophylls, two bac- teriopheophytins and two quinones.Two bacteriochlorophylls, BChl,, form the primary donor and one bacteriopheophytin serves as the primary acceptor. E.s.e. determination of the anisotropic 15N hyperfine interactions have shown that the primary donor cation resides symmetrically in two of the four BChls in the form of a special pair of bacterial chlorophylls, BChlz. In direct contrast, our e.s.r. studies on the triplet state of the primary donor in single crystals of reaction centres suggest that the triplet state resides asymmetrically in BChl,, where the triplet is highly asymmetrical in R. viridis but is much more symmetrical (C,) in R.sphaeroides. RYDMR studies of the initial radical pair formed indicate that the extra bacteriochlorophyll molecule, BChl,, that is between the special pair donor cation and the primary acceptor bacteriopheophytin anion is not involved in a discrete electron-transfer step in bacterial reaction centres. By combining the results of our magnetic-resonance experiments with X-ray structural information, the following description emerges: (1) the ground-state, primary donor is a supermolecule dimer with approximately C, symmetry; (2) the bridging Bchl, molecule probably functions as a superexchange site for rapid transfer of electrons from the primary donor to the primary acceptor but with negligible back reaction; (3) the special pair, BChl,, is lower in energy than the bridging molecule, BChl,, such that the initial radical pair is formed via super-exchange between the distant (10 A edge-to-edge) special pair BChl, and the bacteriopheophytin.14 Magnetic Resonance of Ultrafast Chemical Reactions The initial act of photosynthesis is the formation of a radical pair BiP-.The purpose of this paper is to characterize this initial radical pair using magnetic techniques. The primary donor (B,) is composed of a pair of bacteriochlorophylls.l Since B, is part of the radical pair, the nature of B, is involved in the present investigation. Whether B, acts as a monomer or a dimer is important to any characterization of the initial radical pair. The initial acceptor (P), a bacteriopheophytin,2 is an accepted monomer anion in the initial radical-pair state.QFe is a quinone-iron complex3 that serves as a secondary electron acceptor. In addition, a bacteriochlorophyll (not shown here) may be a bridging molecule between B, and P.4 Such a bridging intermediate may be pertinent to the radical-pair state. The electron transfer from *B, to P to form the initial radical pair takes place in < 6 P S . ~ The secondary electron acceptors permit additional electron transfer to proceed from P- to Q in ca. 200 P S . ~ ~ If Q is removed or previously saturated by electrons 'in the dark' then charge annihilation and recombination occurs on a 10 ns timescale5 to form a triplet 3*B2. Otherwise the fast forward reaction to form Q- proceeds. As a consequence of the relatively slow back reactions the efficiency of the primary reaction in the photo-oxidation of BZ exceeds 98% .8 y An important feature of this radical-pair reaction, which distinguishes it from ordinary in vitro liquid solution reactions, is that it takes place embedded in a protein matrix, essentially in the solid state. Consequently, the participant molecules are maintained in a fixed spatial relationship with continuous electronic interactions among them. The various states shown in the above scheme are often described in terms of discrete loci of electronic charge or excitation energy, by analogy with conventional chemical reactions. The constant proximity of the reactants and strong interactions among them, however, suggest that such a treatment is at best incomplete.A refined description of the primary processes of photosynthesis should include the delocalization of states such as lB?, 3B2*, Bl, and even the radical pair BZP-, over the many components of the reaction centre. Time-resolved magnetic resonance techniques such as electron spin-echo (e.s.e.) and reaction-yield-detected magnetic resonance (RYDMR) provide powerful probes into the nature of the various states of bacterial photosynthesis. In turn, electron spin resonance (e.s.r.) of single crystals of reaction centres provide a means of correlating the properties of these states with the molecular structure of the reaction-centre chromo- phores which has recently been established by X-ray crystallography. lo In fact, much of our current understanding of the primary events of photosynthesis is derived from time-resolved optical spectroscopy and magnetic resonance spectroscopy of reaction centres in solution.Although such techniques afford a detailed description of kinetic events in photosynthesis, only limited structural information is available without the use of single crystals of reaction centres. On the other hand, X-ray diffraction studies on single crystals provide a detailed structural description of the core electrons of the reaction-centre componentslO but no kinetic information. The availability of single crystals of reaction centres for magnetic-resonance studies11-13 provides a powerful and highly accurate means of assigning observed 'working' valence states of the reaction centre unambiguously to its structural features.Although the ground-state special pair model for the primary donor of bacterial photosynthesis has now been confirmed by the X-ray structural data,lo additional work is necessary to establish whether the working states are dimeric in nature. The initial derivation of the special pair model was based on a symmetrical dimer of chlorophyll or bacteriochlorophyll.' The structure of the special pair as presently revealed from the X-ray study is a symmetrical dimer with approximate C, symmetry.lO In this paper we discuss e.s.e. modulation data on 15N-enriched cation of the special-pair donor which confirms the symmetrical nature of the primary-donor cation. In addition to, and in contrast with these data, we also present e.s.r. evidence that the triplet state of the special pair and/or its immediate protein environment deviates substantially and significantly from C, symmetry, and thus has monomer-like properties.J.R . Norris, C. P. Lin and D. E. Budil 15 Finally, we report studies designed to measure the singlet-triplet energy gap, ZJ, in the initial radical pair. This experimental evidence depends on measuring the relative triplet yield (or decay rate of the radical pair) as a function of magnetic-field strength and temperature. In one series of experiments the magnetic field is a static field of strength zero to several hundred gauss. In the other case the magnetic field is a microwave-induced magnetic field in the rotating frame ranging from zero to a few tens of gauss as in a RYDMR experiment.The direct measurement of J places severe limits on the type of intermediates, if any, that might precede the initial radical pair. Experiment a1 Reaction centres of R . sphaeroides (in some cases depleted of quinone) and R. viridis were prepared in the standard manner. R. rubrum, similar to R . sphaeroides in its e.s.r. properties, was studied as chromatophores. The e.s.r., e.s.e. and RYDMR spectrometers have been described previously. The e.s.e. studies of cations were performed on light irradiated samples in which the quinones were not removed or reduced, producing a long-lived cation. The triplets were investigated using light-modulated e.s.r. spectroscopy in single crystals of R. viridis in which the quinone was already in a reduced state in the dark, These single-crystal e.s.r.data have been published previously, but have not been compared with the molecular coordinates of the bacteriochlorophylls as determined by X-ray diffraction. Results and Discussion E.S.E. 15N Anisotropic Hyperfine Coupling Constants for the Primary Donor Cation Although the X-ray structure for the reaction centre of Rhodopseudomonas uiridislO proves the dimeric nature of the ground-state special pair, the first question to be resolved is whether the cation of the special pair behaves like a monomer or like a dimer. The four nitrogens near the centre of the chlorophyll macrocycle can serve as an e.s.e. probe of the cation unpaired electron spin distribution and thus determine the nature of the special pair with respect to the cation state.l* Recent ENDOR studies15 found an average reduction of two in isotropic coupling constants for 15N BChla'+ us.P865+ in R26 R. sphaeroides. However, reduction factors of about five were reported for the anisotropic coupling constants for the same systems.15 The reduction factor of two supports a symmetrical dimer whereas the reduction factor of five does not. Here we argue that the anisotropic 15N coupling constants are the more rigorous parameters for probing the monomeric us. dimeric nature of the primary donor, and thus the reduction factor of five is a serious discrepancy for the cation special pair. The 15N anisotropic hyperfine interaction measures primarily only the nitrogen spin density. In contrast, the isotropic hyperfine interaction of 15N of chlorophylls measures the spin density of the adjacent carbon atoms as well as the nitrogen. Thus the anisotropic interaction is the parameter of choice.To measure this anisotropic quantity we use e.s.e. envelope modulation spectroscopy16~ l7 on 15N-enriched R. rubrum chromatophores. Our e.s.e. results support a set of anisotropic 15N hyperfine couplings for P86Y which have an average reduction factor closer to two, as required by the special pair model. Several stimulated echo modulation curves taken at various z are shown in fig. 1 and 2. The hyperfine coupling constants obtained by a non-linear least-squares fitting procedure of the ratios of different modulation curves are in table 1. In analysing the data we have assumed that the modulation arises from four pairs of approximately parallel, axially symmetric 15N hyperfine tensors.The four isotropic coupling constants are constrained to match the liquid-solution ENDOR mea~urernents,~~ while the four anisotropic coupling constants ( A II components of the traceless, axially symmetric tensors) are varied until the best fit is achieved by the least-squares procedure. A , , and16 Magnetic Resonance of Ultrafast Chemical Reactions c I I 1 I I I I 1 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8 tlw I I 1 I I 1 I I 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8 tlW I Fig. 1. (a) Stimulated echo modulation from the cation of bacteriochlorophyll a. Traces from top to bottom were taken with z setting from 0.28 to 0.56 ps. (6) Calculated modulation using the above t settings and the hyperfine coupling constants (table 1 ) from e.s.e.Each calculated trace is multiplied with an exponential decay constant of 4 ps. The time axis is z+ T. A , (where A , = - A , , / 2 ) denote components of the traceless dipolar hyperfine tensor A . The fitting procedure is analogous to extracting hyperfine coupling constants by fitting powder spectra in the frequency domain, except that the problem of data truncation inherent in all e.s.e. experiments is considerably less severe in the time domain. For more detail, see ref. (18) and (19). The average reduction factor for the four anisotropic coupling constants (relative to BChla+) is 2.3. We therefore conclude that the cation state is a symmetrical dimer special pair in R . rubrum. In addition we argue that the special pair is a symmetrical dimer cation in the radical pair state in R .rubrum and R . sphaeroides .J . R. Norris, C. P . Lin and D . E. Budif 17 I I I 1 1 1 1 1 1.0 2.0 3.0 4.0 5.0 6.0 7.0 a tlP 0.b 1.b 2.b 3.b 4.b 5.b 6.b 7.b 8 tlPS Fig. 2. (a) Stimulated echo modulation from the primary donor cation of R. rubrum. Traces from top to bottom were taken with z setting from 0.28 to 0.56 ps. (b) Calculated modulation using the above z settings and the hyperfine coupling constants (table 1) from e.s.e. Each calculated trace is multiplied with an exponential decay constant of 4 pus. The time axis is z+ T. Single-crystal E.S.R. Studies of Reaction-centre Triplets The initial X-ray structure shows a local, approximate, C, symmetry for the ground state of the electron-transfer components of the photosynthetic reaction centre of R.uiridis including the special pair. A rigorous test of the actual ‘working’ symmetry based on the triplet state of the donor is simple and straightforward. By working symmetry we mean the symmetry exhibited in states other than the ground-state core electrons probed by the X-ray technique, such as the cation state, the lowest excited triplet state or lowest excited singlet.18 Magnetic Resonance of Ultrafast Chemical Reactions Table 1. 15N hyperfine coupling constants for BChla'+ and P865'+ of R. rubrum as determined by e.s.e. modulation BChla'+ P865'+ aiso Aila aiso A I /MHz /MHz /MHz /MHz RFb 3.16 2.43 1.0Y 0.57 4.26 3.41 2.61 1.61 1.34 1.95 4.10 2.67 2.22 1.77 1.51 4.45 2.91 2.60 1.96 1.48 a A l l denotes the largest element of the traceless and axially symmetric dipolar tensor.R , is defined as the ratio A lI(BChla'+)/A ,,(P865'+). Isotrophic coupling constants in this column are taken from liquid-solution ENDOR measurements.* The triplet-state method used to probe working symmetry involves comparing the principal directions for the zero-field splitting tensor of the triplet state (in a single crysta1)llY l2 to the direction of the C, axis as determined by X-ray diffraction.1° If the triplet state resides symmetrically in a special pair of bacteriochlorophylls with C, symmetry then one of the three triplet axes, either x, y or z , must lie along, or approximately along, the two-fold rotation axis as defined by the C, symmetry. Of course the two remaining triplet-state axes must then be perpendicular to the C, axis. Since the X-ray crystal structure has not yet been refined, the C, axis direction is not expected to be perfect.Also the symmetry of the immediate protein environment as yet has not been reported by X-ray diffraction techniques. For the purposes of spectroscopic observations, the environment is as important as the geometry of the bacteriochlorophyll molecules in the symmetry/asymmetry question. The orientation of the zero-field axes of the molecular triplet in single crystals of both R . viridis and R. sphaeroides has been detennined.ll9 l2 E.s.r. analysis of single crystals in the triplet state show that no zero-field triplet axis is approximately parallel to the X-ray determined C, axis. To the contrary, the total angle between the C, and the triplet z-axis is closer to 45" than O 0 ., O The reported zero-field axes belong to one of the eight possible triplet orientations in the unit cell; therefore, we compared each site to the reported direction of the C, axis.l* Table 2 shows the directional angles of the C, axis with the zero-field axes for each of the eight orientations of the P890 triplet in the unit cell. Based on this tabulation no zero-field triplet axis is even approximately parallel (or antiparallel) to the C, axis at any orientation. The closest approach is 27" (considering only a single direction cosine), much larger than possible errors in the direction of the triplet or the C, axes. Thus, the triplet state of the P890 special pair deviates significantly from C, symmetry, despite the highly symmetric structure indicated by the X-ray data.Such asymmetry probably reflects an asymmetry in the protein environment immediately surrounding the triplet state of the special pair donor. The observed triplet axes may be more closely related to those of a monomeric triplet excited state residing on one of the BChl b molecules. In order to make this comparison, approximate zero-field axes for each BChl were derived using the atomic coordinates from the crystal structure kindly provided by Deisenhofer and Michel. The z axis wasJ . R. Norris, C . P . Lin and D. E. Budil 19 Table 2. Directional angles (in degrees) of the C, axis in the molecular coordinates for the eight triplet orientations in the unit cell of R. viridis site X Y z 27 138 100 142 118 71 54 69 116 48 148 128 68 127 39 47 95 89 60 94 37 137 78 129 taken as the normal vector to a plane fitted by least-squares to the coordinates of the members of the conjugated ring; the y direction was assigned to the projection of the ni trogen-t o-ni trogen vector on to the plane.Table 3 gives the direction angles of the measured triplet axes with the assumed monomer axes of the special pair BChls for the eight crystal sites. Only one BChl of the special pair in the unit cell coincides appreciably with the experimental axes (L subunit, site l), but the agreement for this case is excellent: the three measured axes are all within 5" of the assumed monomer axes. We are using the standard nomenclature of L (low molecular weight) and M (medium molecular weight) to denote the two protein subunits that contain the chemically active pigment molecules and quinones.The L protein is the only photoactive subunit. The P890 triplet appears to reside primarily on the L-subunit side of the dimer, which is the side closer to the menaquinone site revealed in the X-ray data. Fig. 3 depicts the experimental axes and the assumed monomer axes relative to the molecular framework of the special pair. The P890 triplet is not simply a BChl b monomer triplet, since its zero-field splitting parameters D and E are smaller than the observed values for in vitro BChl b monomer triplets. One way to reduce D and E is to redistribute the electrons by including charge-transfer character in the P890 triplet. We approximate D = -61 G for a pure charge-transfer triplet [BChl(L) + BChl(M) -3 by distributing the unpaired spins over the atomic coordinates according calculated spin densities for BChl b ions2' Ca.23% charge-transfer character is required to reduce D from the monomeric value of 227 G to the in vivo value of 165 G. Thus by including triplet-state charge transfer within the special pair the directions of the triplet zero-field tensor directions agree with the three measured directions within ca. 3", as shown in table 4. Dimers of organic molecules often display dimeric optical properties for the lowest excited singlet state but monomeric properties for the triplet state.22 In these cases the triplet state is 'self-trapping' and as a result resides on both 'monomer' halves of the dimer but not at the same time.The rationale for a given state behaving as monomer or as dimer is based primarily on the differences in spin angular momentum for the state involved. For example, the spin angular momentum restrictions for sharing a state or excitation between two monomers is rather severe for triplets. However, for singlets or cations spin restrictions are much less important. Thus the triplet state provides a severe test of the symmetry of the special pair and its local environment. We emphasize that the triplet of R . viridis correlates only with one half of the special pair. If the triplet state itself were responsible for the asymmetry, i.e. the triplet were self-trapping, then the triplet would sometimes be found on the M side as well as the L side. Since we find the triplet axes to correlate only with the L side, we conclude that the observed asymmetry20 Magnetic Resonance of Ultrafast Chemical Reactions Fig.3. Bacteriochlorophyll special pair in R. viridis crystal. The long arrows locate the L- bacteriochlorophyll 'monomer' x, y and z. The short arrows are the observed x, y and z axes of the triplet state. By allowing 23 % charge transfer between the L-bacteriochlorophyll of the special pair and the M-bacteriochlorophyll the 'monomer' arrows and the observed arrows all coincide within 3". of triplet distribution probably reflects an asymmetry in the protein environment immediately surrounding the special pair. Such an asymmetry in the protein environment of the special pair also suggests that the cation of the special pair of R .viridis is probably asymmetrical. The observed e.s.r. linewidth of the cation (ca. 11.5 G ) is too broad to indicate a symmetrical dimer and too narrow to indicate a monomer. Thus the cation e.s.r. linewidth also suggests an asymmetrical spin distribution between the two chlorophylls of the special pair of R. viridis. However, the much narrower line widths of ca. 9.5 G for the cation of R. sphaeroides or R . rubrum suggests that the cation state in these organisms is much more symmetrical, in agreement with the anisotropic 15N hyperfine coupling constants. We take the triplet evidence to support the view that the exact nature of the special pair ranges from a highly symmetrical dimer to an asymmetrical dimer. Even in the asymmetrical case the special pair is acting essentially as a supermolecule, since the triplet zero-field constants are reduced relative to a monomer of bacteriochlorophyll.In addition the triplet asymmetry suggests that the pigments of the M protein subunit, except for the special pair, are not very important to the characterization of the initial radical pair. Static and Timedependent Magnetic-field Effects in Bacterial Photosynthesis The effect of an applied magnetic field upon the quantum yield of triplets produced in photosynthetic reaction centres of bacteria has been widely studied to gain insight intoJ . R . Norris, C . P . Lin and D. E. Budil 21 the dynamics and energetics which govern the primary stages of photoinduced charge separation. Of particular interest are the very weak electronic exchange interactions which can be measured by magnetic-field effects, for these may be related to the rates of electron transfer within the reaction centre.To date, magnetic exchange in the primary radical pair has been determined indirectly by a parametric fit to the observed low field dependen~e,~-,~ or resonant microwave power dependence26-28 of the triplet yield, or by fitting the measured electron spin polarization in Fe-depleted reaction centres. 29 We report here the first direct measurement of isotropic electron exchange in quinone- depleted reaction centres from the bacterium R. sphaeroides R-26. Triplets arise in the reaction centre by charge recombination between the primary acceptor anion (P.-) and the primary donor cation (B;+) when electron transfer to a secondary acceptor is blocked.The mechanism for 3B2 formation is analogous to the nuclear hyperfine-induced intersystem crossing found in reactions of radicals in solution;30* 31 however, since the photosynthetic reaction takes place embedded in the reaction-centre protein, electron exchange is not modulated by diffusion of the radicals as it is in liquid solution. Effects of static external magnetic fields on 3B, yield are usually explained in terms of the energy levels of the spin states of the Bi+P'- ion pair. The radical pair is born in the singlet state, which is non-stationary because of nuclear hyperfine terms which mix it with the triplet states. At zero field, for sufficiently small singlet-triplet energy gaps, transitions can occur to all three of the triplet sublevels.When a static magnetic field is applied, two of the triplet states are removed by the Zeeman energy, and transitions occur only to one of the sublevels, resulting in a reduction of the 3B2 yield measured upon charge recombination. An isotropic exchange interaction in the radical pair gives a singlet-triplet energy gap E,- Et = 2J, which results in a singlet-triplet level crossing at an applied external field gPBo = 124. At this point, radical-pair intersystem crossing is most efficient, and therefore an initial increase in triplet yield from B, = 0 to B, = 124/gP could be expected if isotropic exchange is present. A similar effect occurs when a microwave field is applied in resonance with the energy differences in the radical pair.The analogy between the static and resonant magnetic field experiments becomes clear when the radical pair spin states are viewed in the reference frame rotating at the microwave frequency, with z defined along the direction of the microwave field B,. The magnitude of the B, field is varied from zero to a few tens of gauss by adjusting the power of the microwaves. Since the B, field is stationary in this reference frame, the experiment is entirely analogous to turning on a static field in the laboratory frame, although the effective Hamiltonian of the system is changed by the transformation into the rotating-z frame. Again a maximum in relative triplet yield should occur when B, = 2J,279 28 in the absence of significant homogeneous or inhomo- geneous broadening terms.An exchange resonance does appear in the static field effect on reactions between radicals in solution which are linked to each other by a methylene chain, as observed either optically32 or by the CIDNP effect.33 Some theoretical calculations of the field effect on reaction centres predict an observable peak under certain c i r c ~ m ~ t a n c e ~ . ~ * - ~ ~ No resonance has yet been reported in experiments on quinone-depleted reaction centres from several laboratories at 0 0C23 and room temperat~re.~~ Some workers have concluded from this that 2J is zero or negligible in the radical ~air;,~g 35 however, most recent model calculations include J as a parameter to fit the observed triplet-yield field dependence. When the temperature is dropped below 0 "C, a peak does appear at lBol = 14 G in the low-field profile of reaction-centre triplet yield, as shown in fig.4. The resonance is observable at the absorption of 3B, at 420 nm, the ground-state B, absorption band at 870 nm, and in plots of radical-pair lifetime us. 38 Although the amplitude of the peak relative to zero field varies slightly among different samples, the resonance has22 Magnetic Resonance of Ultrafast Chemical Reactions Table 3. Overlap matrices for the measured x, y and z directions of the triplet state2 and the molecular coordinate x, y and z directions determined from X-ray diffraction data of R. viridis'O BCMP BCLP BCMP BCLP e.s.r. axis X Y Z X Y Z X Y Z X Y Z X Y z X Y z X Y z X Y z 165 100 101 102 15 81 80 79 165 130 87 40 98 172 92 41 97 50 109 32 115 20 76 104 96 119 151 77 31 63 164 83 75 81 120 32 3 87 90 51 80 139 118 99 30 93 5 87 75 165 90 91 169 101 90 93 3 137 101 131 28 97 62 64 68 145 12 93 79 164 106 93 69 155 103 85 8 97 106 27 69 146 101 122 101 83 13 93 69 158 78 37 56 91 148 122 72 154 72 154 94 64 152 104 67 36 85 125 67 127 45 118 61 138 61 65 41 120 33 101 83 149 60 91 138 132 33 65 111 25 97 114 139 116 60 103 110 156 66 60 40 131 59 123 0.040 0.055 !4 8 d- n 0.030 c, 0 3 -2 8 2 Q 0.02s W B' 0.020 -50 I I 1 1 50 100 150 200 magnetic field/(; Fig.4. Triplet yield in quinone-depleted reaction centres from R. sphaeroides R-26 as a function of applied external magnetic field near 0 G. Triplet absorbance was measured by a least-squares fit of an asymptote to radical pair decay measured at 420 nm up to ca.180 ns after excitation with 600 nm light. Conditions: 42 pmol dm-3 reaction centres in buffer containing 60% ethylene glycol, 0.01% LDAO in a 0.25 mm path at 243 K. appeared consistently at lBol = 14 G in several different reaction-centre preparations, independent of the detergent (LDAO or BRIJ-58) and low-temperature glass (glycerol, sucrose or ethylene glycol) used. The position of the resonance is also independent of the direction of polarization of excitation light at 600 nm relative to the field direction (data not shown). We therefore ascribe the peak to an isotropic electron exchange interaction 124 = 14 G in the primary radical pair in the bacterial reaction centre.J. R. Norris, C.P. Lin and D. E. Budil 23 1.10 1.05 5 9" 0 1 r4 d Y m g 0.95 E .- t o - 0.90 0.85 0 \ * I I I I I 1 5 10 15 20 25 30 B , / G Fig. 5. Relative triplet yield as a function of microwave B, with triplet yield determined at 420 nm. Conditions: 80 pmol dmP3 reaction centres in buffer containing 0.01 % LDAO in 0.1 mm path at room temperature. Line: Theoretical triplet yield calculated according to ref. (25) with z, = 25 ns (singlet radical-pair lifetime) and rt = 2 ns (triplet radical-pair lifetime). The (24 from the static-field study agrees closely with the value of 16 G recently estimated from initial RYDMR spectroscopy of the radical pair.26 Fig. 5 shows a room-temperature field study taken in the rotating frame which exhibits a resonance at B, = 13 2 G, again demonstrating the excellent agreement of RYDMR, static magnetic-field measurements and theoretical calculations of the microwave field effect.27 For this experiment the B, field was carefully calibrated against applied power by measuring the duration of 7r/2 and 71 pulses in our e.s.e. spectrometer. It is noteworthy that the J resonance appears in the rotating-z spectrum but is obscure in the static-field profile at room temperature. As Lersch and Michel-Beyerle have rioted,28 a major difference between the high- and low-field spin Hamiltonians is the presence of electronic magnetic dipolar terms mixing the m, = 1 triplet levels at low field, which are neglected in the high-field approximation used to describe RYDMR. This provides some indication that the dipolar interaction D in the primary radical pair may not be neglected.However, the obscurity of the resonance in the static-field dependence and the relatively high temperature at which it appears permits strict limits to be placed on D and the kinetic parameters of the radical pair. The temperature at which the peaks are first discernible in the static-field study varies with detection method. The signal-to-noise ratio is generally best in the lifetime spectrum, since the fitting of a single exponential to the observed radical pair decay acts as a noise filter. Fig. 6 illustrates the appearance of the peaks in the lifetime spectrum of the radical pair with decreasing temperature; they are faintly visible at temperatures as high as 0 "C, and reach their maximum amplitude by ca.- 30 "C. The peaks persist at 14 G down to ca. 130 K, at which point they are no longer visible above the noise because of the diminishing relative field effect as the triplet yield approaches unity. If the uncertainty broadening of the triplet levels is to be < 14 G, the decay processes must have characteristic times z > 2 ns below - 30 0C.39 The limits which may be placed on the magnetic dipolar interaction D depend upon the relative signs of D and J . Making the reasonable assumption that D < 0 for a radical pair, if 2J > 0 (singlet higher in24 Magnetic Resonance of Ultrafast Chemical Reactions Table 4. Overlap matrices comparing charge-transfer triplet states in the special pair of R. viridisa calculated axis fraction e.s.r. charge axis x Y z transfer DIG E/G 0.00 227 59 X 3.4 86.6 90.0 y 93.4 4.8 86.7 z 89.8 93.3 3.3 X 4.1 86.0 91.0 y 93.9 4.7 87.3 z 88.8 92.6 2.8 X 4.7 85.7 91.7 y 94.3 4.8 87.8 z 88.1 92.1 2.8 X 4.8 85.6 91.9 y 94.4 4.8 87.9 z 87.9 91.9 2.8 0.15 187 51 0.23 166 46 0.25 160 45 a The L-bacteriochlorophyll ‘monomer’ contains most of the triplet state.The M-bacteriochlorophyll ‘monomer’ of the special pair participates in the triplet state of the primary donor via a small amount of charge transfer. energy) then 12D/31 should be < 124 for the singlet state to lie outside the dipolar width, i.e. ID1 < 21 G. In the case 2 J < 0, the appearance of the J-resonance suggests [Dl < 42 G, although this limit is not as accurate. An estimate of the sign of J is available from model calculations of the RYDMR spectrum of the radical pair,26 which indicate that the singlet radical pair state is higher in energy.Recent calculations by Hoff and H ~ r e ~ ~ of electron spin polarization in Fe-depleted reaction centres support this assignment. Hoff and Hore actually assert that their determination of the sign of J conflicts with the earlier RYDMR result; however, these authors did not take into account the difference between ref. (26) and (29) in sign convention for the exchange interaction. In fact, the two results are consistent, indicating a ferromagnetic (singlet higher) exchange in the radical pair. If these determinations are correct, they would indicate the more severe restriction on the value of D in the radical pair, which is not consistent with the values for D obtained from RYDMR26 and some other field-effect 3 4 9 36 However, such a limit is compatible with the value for D in the radical pair in reaction centres of R.viridis which has been calculated using the atomic coordinates of the B, and P radicals recently available from X-ray diffraction A significant aspect of the measured exchange term is that it is far too small to account for the rapid rates of forward electron transfer in the reaction centre.41 One possible explanation for this discrepancy is that 2J arises from a kinetic hopping between a ‘distant’ radical pair with no exchange and a ‘close’ radical pair with significant electronic interaction^.^' This model is consistent with a two-step charge separation, with the electron first transferred to the intermediate BChl and then to P.42 Alternatively, the primary charge separation might be accomplished via a superexchange 44 in this case the radical pair would be better described as a coherent rather than a kineticJ.R. Norris, C. P. Lin and D. E. Budil 30 - 25 - 2 ; 20- \ n .ii 1 W - 15 - 25 222 K U 2 K 251 K 272 K 296 K 10 : I I I 1 I 1 -100 -50 0 50 100 150 200 magnetic field/(; Fig. 6. Radical-pair lifetime as a function of applied static magnetic field and temperature. Lifetimes were determined by least-squares fit of a single exponential to radical-pair decay measured at 420 nm. Conditions the same as for fig. 4. combination of electronic configurations. Such a mechanism can reconcile a small spin exchange with fast electron transfer because the charge-transfer terms contributing to the radical pair singlet-triplet energy gap enter into the calculation at much higher order for superexchange than for direct The ability to measure spin exchange in the radical pair directly may provide a sensitive test of proposed mechanisms for primary charge separation in photosynthesis.The observed temperature independence of 124 suggests that spin exchange does not arise from a kinetic mechanism; our results are more consistent with a description of the BZP- radical pair as a coherent combination of supermolecular electronic configuration^.^^ Summary In conclusion, we wish to emphasize that the relatively high temperature at which the J-resonance can be detected permits a substantial refinement of the parameters used to characterize electronic interactions'in the reaction centre at physiological temperatures.In addition, the possibility of measuring J directly provides an important experimental means of testing proposed mechanisms for the origin of magnetic exchange in the primary radical pair. A more complete treatment of the temperature dependence of magnetic exchange in bacterial reaction centres is in preparation. Because of the small J value and the high magnetic resolution (of the order of J) of these radical pair experiments, we can rule out any significant kinetic contribution to spin exchange in the radical-pair state. In other words, these magnetic-field investigations26 Magnetic Resonance of Ultrafast Chemical Reactions of the initial radical pair suggest that the extra bacteriochlorophyll molecule, BChl,, that is between the special-pair donor cation and the primary acceptor bacteriopheophytin anion is not involved in a discrete electron-transfer step in bacterial reaction centres.By combining the results of our magnetic-resonance experiments with X-ray structural information, the following description emerges: (1) the ground state, the cation state and the triplet state of the primary donor are each a supermolecule dimer with approximately Cz symmetry; (2) the bridging BChl, molecule probably functions as a superexchange site for rapid transfer of electrons from the primary donor to the primary acceptor but with negligible back reaction; (3) the special pair, BChl,, is lower in energy than the bridging molecule, BChl,, such that the initial radical pair is formed via super exchange between the distant (10 A edge-to-edge) special pair BChl, and the bacteriopheophytin and (4) the triplet is a sensitive probe of the asymmetry of the reaction centre.We thank Dr J. Deisenhofer and Dr H. Michel for providing X-ray data coordinates from R . viridis. We also thank Mr S. Kolaczkowski for help in measuring the relative triplet yield as a function of B,. This work was supported by the Division of Chemical Sciences, Office of Basic Energy Sciences, U.S. Department of Energy, under Contract W-31-109-Eng-38. References 1 J. R. Norris, R. A. Uphaus, H. L. Crespi and J. J. Katz, Proc. Nut1 Acad. Sci. USA, 1971, 68, 625. 2 J. Fajer, D. C. Brune, M. S. Davis, A. Forman and L.D. Spaulding, Proc. Natl Acad. Sci. USA, 1975, 3 G. Feher, M. Y. Okamura and J. D. McElroy, Biochim. Biophys. Acta, 1972,267,222. 4 C. Kirmaier, D. Holten and W. W. Parson, FEBS Lett., 1985, 185, 76. 5 W. W. Parson, R. K. Clayton and R. J. Cogdell, Biochim. Biophys. Acta, 1975, 387, 265. 6 K. Kaufmann, P. L. Dutton, T. L. Netzel, J. S. Leigh and P. M. Rentzepis, Science, 1975, 188, 1301. 7 M. B. Rockley, M. W. Windsor, R. J. Cogdell and W. W. Parson, Proc. Natl Acad. Sci. USA, 1975,72, 8 C. A. Wraight and R. K. Clayton, Biochim. Biophys. Acta, 1974, 333, 246. 9 C. A. Wraight, J. S. Leigh, P. L. Dutton and R. K. Clayton, Biochim. Biophys. Acta, 1974, 333,401. 72, 4956. 225 1. 10 J. Deisenhofer, 0. Kepp, K. Miki, R. Huber and H. Michel, J. Mol. Biol., 1984, 180, 385.1 1 P. Gast, M. R. Wasielewski, M. Schiffer and J. R. Norris, Nature (London), 1983, 305, 451. 12 P. Gast and J. R. Norris, FEBS Lett., 1984, 177, 277. 13 J. P. Allen, R. Theiler and G. Feher, Biophys. J., 1985, 47, 4a. 14 P. J. O’Malley and G. T. Babcock, Advances in Photosynthesis Research, ed. C. Sybesma (Martinus 15 W. Lubitz, R. A. Isaacson, E. C. Abresch and G. Feher, Proc. Natl Acad. Sci. USA, 1985, 81, 7792. 16 L. G. Rowan, E. L. Hahn and W. B. Mims, Phys. Rev. A, 1965, 137, 61. 17 W. B. Mims, Phys. Rev. B, 1972, 5, 2409. 18 C. P. Lin and J. R. Norris, to be published. 19 J. Tang, C. P. Lin and J. R. Norris, J. Chem. Phys., 1985, 83, 4917. 20 J. R. Norris, D. E. Budil, H. L. Crespi, M. K. Bowman, P. Gast, C. P. Lin, C. H. Chang and M. Schiffer, Antennas and Reaction Centers of Photosynthetic Bacteria, ed.M. E. Michel-Beyerle (Springer- Verlag, Berlin, 1985). 21 M. S. Davis, A. Forman, L. K. Hanson, J. P. Thornber and J. Fajer, J. Chem. Phys., 1979, 83, 3325. 22 J. B. Birk, Photophysics of Aromatic Molecules (Wiley Interscience, New York, 1970). 23 A. Ogrodnik, H. W. Kruger, H. Orthuber, R. Haberkorn and M. E. Michel-Beyerle, Biophys. J., 1982, 24 C. E. D. Chidsey, C. Kirmaier, D. Holten and S. G. Boxer, Biochim. Biophys. Acta, 1984, 766, 424. 25 J. Tang and J. R. Norris, Chem. Phys. Lett., 1982,92, 136. 26 J. R. Norris, M. K. Bowman, D. E. Budil, J. Tang, C. A. Wraight and G. L. Closs, Proc. Natl Acad. 27 J. Tang and J. R. Norris, Chem. Phys. Lett., 1983, 94, 77. 28 W. Lersch and M. E. Michel-Beyerle, Chem. Phys., 1983,78, 115. 29 A. J. Hoff and P. J. Hore, Chem. Phys. Lett., 1984, 108, 104. 30 R. Haberkorn and M. E. Michel-Beyerle, FEBS Lett., 1977, 75, 5 . 31 G. L. Closs, J. Am. Chem. SOC., 1969,91, 4552. Nijhoff, The Hague, 1984), vol. I, pp. 697-700. 39, 91. Sci. USA, 1982,79, 5532.J. R. Norris, C. P. Lin and D. E. Budil 27 32 A. Weller, H. Staerk and R. Treichel, Faraday Discuss. Chem. SOC., 1984, 78, 279. 33 G. L. Closs, Chemically Induced Magnetic Polarization, ed. L. T . MUUS, P. W. Atkins and K. A. 34 M. G. Roelofs, C. E. D. Chidsey and S. G. Boxer, Chem. Phys. Lett., 1982, 87, 582. 35 H.-J. Werner, K. Schulten and A. Weller, Biochim. Biophys. Acta, 1978, 502, 255. 36 A. J. Hoff, Q. Rev. Biophys., 1981, 14, 599. 37 M. R. Wasielewski, C. H. Bock, M. K. Bowman and J. R. Norris, J. Am. Chem. Sac., 1983,105,2903. 38 M. R. Wasielewski, C. H. Bock, M. K. Bowman and J. R. Norris, Nature (London), 1983, 303, 520. 39 M. E. Michel-Beyerle, H. Scheer, H. Seidlitz, D. Tempus and R. Haberkorn, FEBS Lett., 1979,100,9. 40 A. Ogrodnik, W. Lersch, M. E. Michel-Beyerle, J. Deisenhofer and H. Michel, Antennas and Reaction Centers of Photosynthetic Bacteria, ed. M. E. Michel-Beyerle (Springer-Verlag, Berlin, 1985). 41 R. Haberkorn, M. E. Michel-Beyerle and R. A. Marcus, Proc. Nut1 Acad. Sci. USA, 1979, 76, 4185. 42 V. A. Shuvalov and W. W. Parson, Proc. Nut1 Acad. Sci. USA, 1981,78, 957. 43 J. R. Miller and R. V. Beitz, J. Chem. Phys., 1981, 74, 6746. 44 A. M. Kuznetsov and J. Ulstrup, J. Chem. Phys., 1981,75,2047. 45 J . Yamashita and J. Kondo, Phys. Rev., 1958, 109, 730. 46 F. Keffer and T. Oguchi, Phys. Rev., 1959, 115, 1428. 47 P. Bertrand, Chem. Phys. Lett., 1985, 113, 104. McLauchlan (Reidel, Boston, 1977). Paper 6/ 1000; Received 22nd May, 1986